Assessing the Precision of Machine Learning Algorithms in Diabetes Mellitus Classification

In order to effectively treat Type 1 diabetes mellitus (T1DM), blood glucose control is essential. Artificial pancreas and insulin bolus calculators, which use machine learning to predict subcutaneous glucose levels, are examples of contemporary digital therapeutic techniques for T1DM patients. Glycemic control is improved by this machine learning application. In the field of healthcare and medical research, deep learning—a subset of machine learning—has gained popularity due to its versatility and ability to perform a variety of tasks well, including illness identification and patient health status prediction. The research effort involves the creation and assessment of Random Forest, Gaussian Naive Bayes, with the shared objective of increasing accuracy and precision in the classification of diabetes mellitus in humans. Both accuracy and precision-wise, Random Forest (RF) outperforms Naive Bayes (NB). While NB falls short with accuracy and precision of 78% and 84%, respectively, RF has exceptional accuracy and precision of 91% and 92%.